Profilin-1 expression is associated with high grade and stage and decreased disease-free survival in renal cell carcinoma

Human Pathology (2014) xx, xxx–xxx

www.elsevier.com/locate/humpath

Original contribution

Profilin-1 expression is associated with high grade and stage and decreased disease-free survival in renal cell carcinoma☆,☆☆ Jason R. Karamchandani MD a,b , Manal Y. Gabril MD, FRCPC c , Rania Ibrahim MD b , Andreas Scorilas PhD d , Emily Filter MD, FRCPC c , Antonio Finelli MD e , Jason Y. Lee MD, MHPE, FRCSC f , Michael Ordon MD, FRCSC f , Maria Pasic PhD a,g , Alexander D. Romaschin PhD a,b , George M. Yousef MD, PhD, FRCPC a,b,⁎ a

Department of Laboratory Medicine and Pathobiology, University of Toronto, Toronto, Ontario, Canada M5S 1A8 Department of Laboratory Medicine and the Keenan Research Centre for Biomedical Science at the Li KaShing Knowledge Institute of St. Michael’s Hospital, Toronto, Ontario, Canada M5B 1T8 c London Health Sciences, London, Ontario, Canada N6A 5A5 d University of Athens, Athens, Greece 106 79 e Division of Urologic Oncology, Princess Margaret Hospital, University Health Network, Department of Surgery, University of Toronto, Toronto, Ontario, Canada M5G 2M9 f Division of Urology, St. Michael’s Hospital, Toronto, Ontario, Canada M5B 1W8 g Department of Laboratory Medicine, St. Joseph’s Health Centre, Toronto, Ontario, Canada M6R 1B5 b

Received 28 August 2014; revised 13 November 2014; accepted 14 November 2014

Keywords: Renal cell carcinoma; Prognosis; Personalized medicine; Tumor markers; Profilin; Pfn1; Metastasis

Summary Clear cell renal cell carcinoma (ccRCC) is associated with high mortality, although individual outcomes are highly variable. Identification of patients with increased risk of disease progression can guide customizing management plan according to disease severity. Profilin-1 (Pfn1) has been recently identified as overexpressed in metastatic ccRCC compared with primary tumors. We examined Pfn1 expression in a tissue microarray of 384 cases of histologically confirmed primary ccRCC with detailed clinical follow-up. Profilin-1 expression showed both cytoplasmic and nuclear staining patterns. The immunoexpression of Pfn1 was scored in a semiquantitative fashion. There was no significant difference in Pfn1 expression between normal kidney and kidney ccRCC. Our results show that strong cytoplasmic Pfn1 expression is associated with high-grade (P b .001) and high-stage (III-IV) (P = .018) disease. Univariate analysis of the data set showed that higher Pfn1 expression is associated with significantly shorter disease-free survival (hazard ratio 7.36, P = .047) and also lower overall survival. Kaplan-Meier analysis showed that high cytoplasmic expression of Pfn1 was also associated with a statistically

Abbreviations ccRCC, Clear cell renal cell carcinoma; Pfn1, Profilin 1; OS, Overall survival; DFS, Disease-free survival. Competing interests: The authors have nothing to disclose nor do they have any conflicts of interest. ☆☆ Funding/Support: This work was supported by grants from the Canadian Institute of Health Research (MOP 119606) (Ottawa Ontario, Canada), Kidney Foundation of Canada (KFOC130030) (Toronto Ontario, Canada), Kidney Cancer Research Network of Canada, and Prostate Cancer Canada Movember Discovery Grants (D2013-39). ⁎ Corresponding author at: George M. Yousef, MD, PhD, FRCPC (Path), MSc, MBBCh, Department of Laboratory Medicine, St. Michael's Hospital, 30 Bond Street, Toronto, ON, Canada M5B 1W8. ☆

http://dx.doi.org/10.1016/j.humpath.2014.11.007 0046-8177/© 2014 Elsevier Inc. All rights reserved.

2

J. R. Karamchandani et al. significant lower disease-free survival (P = .018). It was also associated with lower overall survival, although this was not statistically significant. Profilin-1 lost its prognostic significance in the multivariate analysis when controlling for grade and stage. Profilin-1 expression was not associated with significant prognostic deference in the subgroup of patients with stage 1 disease. Our results suggest that the evaluation of Pfn1 by immunohistochemistry may help to identify patients with an increased risk of disease progression. We validated our results at the messenger RNA level on an independent patient cohort. Higher messenger RNA expression of Pfn1 is associated with significantly lower survival. © 2014 Elsevier Inc. All rights reserved.

1. Introduction Clear cell renal cell carcinoma (ccRCC) is the most common primary renal malignancy occurring in adults [1]. This tumor is associated with significant morbidity and mortality but, nonetheless, has a wide variance in individual patient outcomes—despite a reasonably uniform histologic appearance as compared with other malignancies with heterogeneous morphology and relatively uniform outcomes such as glioblastoma multiforme [2]. Until relatively recently, treating oncologists lacked useful therapeutic agents that demonstrated efficacy in patients with metastatic ccRCC, until the advent of sunitinib therapy [3] coupled with the recent development of several new second-generation small molecule kinase inhibitors approved for clinical use [4]. The need for pathologists to help identify patients with ccRCC at increased risk of disease progression and risk for metastatic spread is urgent because improved prognostic information now has the potential to better inform and optimize screening protocols and therapeutic regimens, heralding a new era of personalized medicine in kidney cancer [5]. The introduction of molecular profiling approaches that allow for simultaneous analysis of thousands of molecules [6] has significantly enhanced biomarker discovery, and independent additional validations are already showing the utility of leveraging a tumor’s molecular signature to inform clinical practice [7,8]. Profilin-1 (Pfn1) is a 140 amino acid protein and major growth regulator of filamentous actin. It participates in many cellular activities, notably the polymerization of actin filaments for which it was originally recognized [9]. When human profilin was cloned in 1988, Northern blot analysis found the greatest concentration in human epithelial, muscle, and renal tissues [10]. Profilin-1 is now well established as a ubiquitously expressed actin monomer–binding protein involved in diverse cellular activities including actin monomer binding, actin polymerization, and transcriptional regulation [10,11]. Profilin-1 is required for cell survival, and double-knockout mouse embryos die before blastocyst formation [12]. Mutations in PFN1 are known to be associated with human disease and have recently been identified as a cause of familial amyotrophic lateral sclerosis [13]. Profilin-1 has recently been implicated in the pathogenesis of several carcinomas, including breast [14] and bladder cancer [15]. Recent results by our

laboratory and others have identified Pfn1 as being compara tively up-regulated in ccRCC as compared with normal kidney [16]. Using quantitative mass spectrometry analysis, we have recently identified several proteins, including Pfn1, that are dysregulated in metastatic as compared with primary ccRCC and demonstrated that they are involved in pathways related to tumor progression and metastasis and thus have the potential of being used as prognostic biomarkers [17,18]. Our current study aimed to evaluate the association between Pfn1 expression and pathologic (grade and stage) and clinical (disease-free survival [DFS] and overall survival [OS]) parameters and its potential utility as a prognostic marker for ccRCC.

2. Materials and methods 2.1. Tissue microarray construction Pure areas from normal kidney cortex tissue and primary ccRCC were selected and circled from donor blocks by a pathologist. Tissue microarray (TMA) blocks containing duplicate 1.0-mm cores of 10% buffered formalin-fixed paraffin-embedded tissue blocks from each specimen were constructed with a manual tissue microarrayer (Beecher Instruments; Sun Prairie, WI). The TMAs contained 384 cases of primary ccRCC obtained from the surgical pathology archives of St. Michael's Hospital between 2001 and 2009. Each block contained 2 marker cores for TMA orientation. Included cases had clinical information that included sex, grade, stage, survival outcomes, and time to disease recurrence if applicable. The study was approved by the research ethics board of St. Michael’s Hospital, Toronto. All cases were primary ccRCC and were reviewed by a pathologist. All new recognized entities including clear cell papillary, translocation carcinomas, etc, were excluded from the analysis. For 80 specimens, matched normal tissues from the same patient were also assessed.

2.2. Immunohistochemistry staining Tissue microarray sections were cut 5-μm thick and placed on charged slides. Slides were deparaffinized in

Profilin-1 is a prognostic marker in renal cell carcinoma

3

xylene, hydrated in gradient ethanol, and pretreated in a microwave oven for 20 minutes at 800 W in 1 L of citrate buffer (0.01 mol/L, pH 6.0) for antigen retrieval. Sections were then incubated with hydrogen peroxide (0.3% vol/vol) in phosphate-buffered saline (PBS) for 15 minutes to quench the endogenous peroxidase activity, followed by blocking with 10% fetal bovine serum in PBS to preclude nonspecific binding. Thereafter, the slides were incubated with Pfn1 primary antibody. Protein expression was detected using the streptavidin-biotin complex with the Dako LSAB+ kit (DakoCytomation; Glostrup, Denmark) and diaminobenzidine as the chromogen. All procedures were carried out at room temperature unless otherwise specified. Slides were washed with 0.025% Triton X 100 in PBS (0.1 mol/L, pH 7.3) (Sigma-Aldrich; Oakville Ontario, Canada) 3 times after each step. Finally, sections were counterstained with Mayer hematoxylin (Sigma-Aldrich; Oakville Ontario, Canada) and mounted with DPX (Sigma-Aldrich; Oakville Ontario, Canada) mounting medium. In the negative control tissue sections, the primary antibody was replaced by isotype-specific nonimmune mouse/rabbit IgG.

Univariate and multivariate analysis was performed to evaluate the hazard ratio (HR), calculated from Cox proportional hazard regression model, for cytoplasmic Pfn1 expression, nuclear Pfn1 expression, grade, and stage with DFS and OS. Multivariate analysis, which adjusted for tumor stage and grade, was also performed. Survival analysis looking at cytoplasmic and nuclear Pfn1 expression was also performed with regard to both DFS and OS.

2.3. Pfn1 interpretation Profilin-1 nuclear staining was scored as positive (N10% of cells staining) or negative (≤10% of cells staining). Cytoplasmic staining was assessed as both percentage staining (proportion of positive tumor cells on the studied section) and staining intensity (intensity of staining by tumor cells), and a combined score was generated. Patients were stratified into either low (≤25% of cells staining, weak staining) or high (26%-100% of cells staining, moderate to strong). Three cases were used as calibrators to define intensity. A total score was obtained by combining both scores. Cores were evaluated independently by 2 pathologists, and controversial cases were assigned a consensus score after discussion. Representative patterns of immunohistochemical staining are shown in Fig. 1. Two cores were evaluated for each patient, and the arithmetic average score was reported. There was an 86% concordance in the expression levels of the 2 cores. Moreover, we compared expression between the TMA cores and whole slide staining in 20 cases. The degree of concordance was 90%.

2.4. Statistical analysis Disease-free survival was defined as the time between the first surgical resection and disease recurrence. Overall survival was defined as the time between the first surgery for primary renal cell carcinoma (RCC) and death for any reason. Tumor grading was done according to the International Society of Urological Pathology Vancouver nucleolar grading system [19]. The association between cytoplasmic and nuclear Pfn1 expression with grade was evaluated using a Pearson χ2 test. The association between cytoplasmic and nuclear Pfn1 expression and low- versus high-stage disease as well as between sexes was evaluated using Fisher exact test.

2.5. Bioinformatics analysis Level 3 gene expression data (normalized gene expression data derived from the Cancer Genome Characterization Center at the University of North Carolina using the RNASeq version 2 RNA-Seq by Expectation Maximization [RSEM] platform) for profilin (Pfn1) messenger RNA (mRNA) in ccRCC and normal kidney and OS data were obtained from The Cancer Genome Atlas, available through the cBio Cancer Genomics Portal (http://www.cbioportal.org/public-portal). Tumors were selected with at least 70% tumor nucleoli concentration and, at most, 20% necrosis, generating a data set of 304 cases. To obtain an optimal cutoff for statistical analysis, we used Cutoff Finder (http://molpath.charite.de/cutoff/) software to generate a prognostic optimal cutoff point to dichotomize Pfn1 mRNA expression as Pfn1 mRNA high expression and Pfn1 mRNA low expression using the log-rank test (P = .0014) [20].

3. Results We first compared the expression of Pfn1 protein in 80 matched normal/cancer specimens from the same patients. The normal kidney cortex was used as an accepted normal counterpart for ccRCC. There was no significant difference in Pfn1 expression between ccRCC and the normal counterpart from the same patient (Table 1). There was a significant positive correlation between Pfn1 cytoplasmic expression and tumor grade with greater frequency of high expression levels in higher tumor grades (P b .001) (Table 1). Higher cytoplasmic expression was also seen in advanced stages (III-IV) compared with earlier stages (I-II) (P = .018). No association was found between Pfn1 expression and sex. The results from the analysis of association between nuclear Pfn1 expression and grade and stage are summarized in Table 2. There was no significant correlation between nuclear Pfn1 expression with either increasing grade (P = .43) or with low versus high stage (P = .57). As with cytoplasmic staining, there was no significantly different expression between male and female patients (P = .19). Univariate analysis (Table 3) showed that higher cytoplasmic Pfn1 expression is associated with a significantly lower DFS (HR 7.36; 95% confidence interval, 1.02-52.91; P = .047). Higher grade and stage were also associated with worse DFS and decreased OS (P b .001).

4

J. R. Karamchandani et al.

Fig. 1 Representative photomicrographs showing the expression of Pfn1 protein by immunohistochemistry. A, Negative nuclear with high cytoplasmic staining. B, Positive nuclear and low cytoplasmic staining. C, Negative nuclear and low cytoplasmic staining. D, Positive nuclear and high cytoplasmic staining. Original magnification ×400.

There was no association between nuclear Pfn1 expression and DFS or OS (Table 3). The HR was not significantly increased in cases with positive nuclear Pfn1 expression for either DFS (P = .131) or OS (P = .470). Multivariate analysis (Table 3) adjusting for tumor stage and grade showed no significant difference in DFS (P = .31) or OS (P = .97) in cases with strong cytoplasmic staining. Similarly, cases with positive nuclear expression

Table 1

also showed no significant increase in the HR for DFS (P = .20) or OS (P = .78) in the multivariate analysis. Kaplan-Meier analysis showed that patients with higher Pfn1 cytoplasmic expression have significantly lower DFS compared with those with lower expression levels (P = .018) (Fig. 2A). No difference was seen in OS (P = .249) (Fig. 2B). Although positive nuclear Pfn1 expression was associated with lower DFS, this did not reach statistical significance

Associations between cytoplasmic profilin expression and clinicopathological variables of cancer patients No. of patients (%)

Variable Status Noncancer Cancer Grade I II III IV Stage I or II III or IV Sex Female Male a b

Total

Pfn1 low expression

Pfn1 high expression

80 362

6 (7.5) 35 (9.7)

74 (92.5) 327 (90.3)

28 178 109 31

7 (25.0) 23 (12.9) 4 (3.7) 0 (0)

21 (75.0) 155 (87.1) 105 (96.3) 31 (100)

211 91

27 (12.8) 4 (4.4)

184 (87.2) 87 (95.6)

.018 b

120 238

13 (10.8) 22 (9.2)

107 (89.2) 216 (90.8)

.38 b

Calculated using Pearson χ2 test. Calculated using Fisher exact test.

P .415 a

b.001 a

Profilin-1 is a prognostic marker in renal cell carcinoma Table 2

5

Associations between nuclear profilin expression and clinicopathological variables of cancer patients No. of patients (%)

Variable Status Noncancer Cancer Grade I II III IV Stage I or II III or IV Sex Female Male a b

Total

Pfn1-negative expression

Pfn1-positive expression

P

80 372

6 (7.5) 18 (4.8)

74 (92.5) 354 (95.2)

.41 a

28 186 109 31

1 7 7 0

(3.6) (3.8) (6.4) (0)

27 179 102 31

(96.4) (96.2) (93.6) (100.0)

.43 b

219 93

11 (5.0) 3 (3.2)

208 (95.0) 90 (96.8)

.57 a

126 242

3 (2.4) 14 (5.8)

123 (97.6) 228 (94.2)

.19 a

Calculated using Fisher exact test. Calculated using Pearson χ2 test.

3.1. Prognostic value of Pfn1 at the mRNA level

(P = .09) (Fig. 3A). There was no association between nuclear Pfn1 expression and OS (Fig. 3B). When patients were stratified by stage, there was no significant association between cytoplasmic Pfn1 expression and either DFS or OS in the subgroup of patients with stage 1 disease (Fig. 4).

Table 3

We also examined the prognostic significance of Pfn1 expression in ccRCC at the mRNA level using the independent The Cancer Genome Atlas data set. Patients were stratified as either Pfn1 high or low expression using an

Cytoplasmic and nuclear profilin expression and patients’ survival DFS

Variable Univariate analysis (n = 352) Cytoplasmic profilin expression Low High Nuclear profilin expression Negative Positive Grade Stage Multivariate analysis c (n = 279) Cytoplasmic profilin expression Low High Grade Stage Nuclear profilin expression Negative Positive Stage

HR

a

OS 95% CI

b

P

HR a

95% CI b

P

1.00 7.36

1.02-52.91

.047

1.00 3.03

0.42-22.1

.273

1.00 4.58 4.27 7.10

0.64-32.89 2.65-6.87 4.29-11.75

.131 b.001 b.001

1.00 2.09 3.78 3.57

0.29-15.19 1.92-7.44 1.75-7.31

.470 b.001 b.001

1.00 2.84 2.6 5.54

0.38-21.24 1.45-4.66 3.17-9.69

.31 .001 b.001

1.00 0.96 2.62 2.71

0.12-7.70 1.11-6.20 1.24-5.90

.97 .03 .01

1.00 3.67 6.97

0.51-26.58 4.20-11.56

.20 b.001

1.00 1.33 3.42

0.18-9.81 1.67-7.05

.78 .001

Abbreviations: DFS, disease-free survival; OS, overall survival; HR, hazard ratio; CI, confidence interval. a Hazard ratio, estimated from Cox proportional hazard regression model. b Confidence interval of the estimated HR. c Multivariate models were adjusted for tumor stage and grade.

6

J. R. Karamchandani et al.

A

A

100

100

Nuclear profilin–negative (n = 16) #Events=1

80

80

60

Survival Probability (%)

Survival Probability (%)

Cytoplasmic profilin– low expression (n = 32) #Events=1

Cytoplasmic profilinhigh expression (n = 313) #Events=79

40

60 Nuclear profilin –positive (n = 339) #Events=83

40

20

20 0 0

0

P = .018

0

12

P = .09

24

48

72

96

120 144 168

24

36

48

60

72

84

96

108 120

DFS Time (Months)

B 100

Nuclear profilin– negative (n = 16) #Events=1

B Cytoplasmic profilin– low expression (n = 31) #Events=1

80

Survival Probability (%)

Survival Probability (%)

100

80

192 216 240

DFS Time (Months)

Cytoplasmic profilin– high expression (n = 321) #Events= 41

60

40

Nuclear profilin-positive (n = 345) #Events=39

60

40

20

20

P = .46

0 0

P = .249

0

24

48

72

96

120 144 168

192 216 240

OS Time (Months) 0

24

48

72

96

120

144

168

192

OS Time (Months)

Fig. 2 Kaplan-Meier curve for DFS (A) and OS (B) of patients with low and high expression of cytoplasmic Pfn1. Patients in the high Pfn1 expression arm have a statistically significant decrease in DFS (P = .018) compared with those in the low Pfn1 expression arm. Patients in the high cytoplasmic Pfn1 arm have a decrease in OS that was not statistically significant (P = .249).

optimized cutoff value (see Materials and methods), and the resulting groups included 193 high-expression patients and 112 low-expression patients. We detected a statistically significant reduction in OS with higher Pfn1 mRNA expression (P = .0014) (Fig. 5).

4. Discussion The extremely strong statistically significant independent association between grade and stage and OS (P b .001) for

Fig. 3 Kaplan-Meier curve for DFS (A) and OS (B) of patients with positive and negative expression of nuclear Pfn1. Patients in the positive nuclear Pfn1 expression arm have a nonstatistically significant decrease in DFS (P = .09) compared with those in the negative nuclear Pfn1 expression arm. There was no association between nuclear Pfn1 expression and OS (P = .46).

this collection of cases is reassuring because any other result would imply that the data set is not representative of the general population. This expected result suggests that it is reasonable to draw conclusions based on this large data set containing 384 cases. Our results have identified Pfn1 as a potential prognostic biomarker in ccRCC. Although the effect was not present in a multivariate analysis after normalizing for grade and stage, in the age of renal biopsies, the utility of a marker that can help identify tumors at increased risk of recurrence is of unquestionable value. The results of this study necessitate a larger study to determine the homogeneity of Pfn1 expression in both large tumors and smaller less than 4 cm tumors (because these are typically more likely to be biopsied).

Profilin-1 is a prognostic marker in renal cell carcinoma

7

A Cytoplasmic profilin– low expression (n = 22) #Events=0

Survival Probability (%)

100

Cytoplasmic profilinhigh expression (n = 147) #Events= 14

80

60

40

20

P = .24

0 0

12

24

36

48

60

72

84

96

108

120

DFS Time (Months)

Fig. 5 Kaplan-Meier curve for OS of patients with high and low expression of Pfn1 mRNA. Similar to the results of immunohistochemical analysis, the curve indicates a statistically significant reduction in OS with higher Pfn1 mRNA expression (P = .0014).

B Cytoplasmic profilin– low expression (n = 21) #Events=1

Survival Probability (%)

100

80

Cytoplasmic profilinhigh expression (n = 153) #Events= 10

60

40

20

P = .76

0 0

12

24

36

48

60

72

84

96 108 120 132 144

OS Time (Months)

Fig. 4 Kaplan-Meier curve for DFS (A) and OS (B) of patients with stage 1 disease with high and low expression of cytoplasmic Pfn1. There was no significant association between cytoplasmic Pfn1 expression and DFS in the subgroup of patients with stage 1 disease (P = .24). There was no significant association between cytoplasmic Pfn1 expression and OS in the subgroup of patients with stage 1 disease (P = .76).

Our current understanding of the behavior of ccRCC is incomplete, and there are currently no tools in clinical use to predict disease outcome apart from the classic clinical parameters. Furthermore, treatment options for primary and metastatic renal cancer are increasing. Accurate data from the pathologic examination of renal cancer specimens will help serve to allow treating clinicians to stratify patients and customize treatment plans for both surveillance and adjuvant therapies, heralding a new era of personalized medicine [21-23]. The identification of markers that can be evaluated by immunohistochemistry is of particular clinical relevance because most laboratories currently have access to immunohistochemical staining, and the test is relatively inexpensive.

Recent evidence has shown the presence of molecular subtypes of ccRCC of distinct prognosis and differential response to targeted therapy. Although these subtypes are indistinguishable by morphology, the use of molecular markers can help to stratify these subtypes with significant clinical implications [24-28]. Our results are in line with recent literature showing that Pfn1 affects cell migration and invasion and tumor aggressive behavior [29]. Reports from other cancers also show the potential utility of Pfn1 as prognostic marker [30] and response to anticancer therapy [31]. Although our study showed no significant variation between normal kidney and kidney cancer overall level of expression, a recent study reported Pfn1 overexpression in RCC [16]. However, this study found that, in normal tissues, tubules stained positive for Pfn1, whereas in RCC tissue, in contrast, the stromal cells in the tumors stained positive. In our analysis, because of restriction of TMA, we were not able to evaluate these components separately with accuracy. An inherent limitation of immunohistochemistry analysis is the suboptimal quantification. In summary, this is the first study evaluating Pfn1 expression in a large data set with rigorous clinical follow-up. We have found that Pfn1 expression is a clinically relevant biomarker in the assessment of ccRCC. Cases with higher cytoplasmic Pfn1 expression have a worse prognosis, whereas cases with lower cytoplasmic Pfn1 expression show better DFS.

References [1] Srigley JR, Delahunt B, Eble JN, et al. The International Society of Urological Pathology (ISUP) Vancouver Classification of renal neoplasia. Am J Surg Pathol 2013;37:1469-89. [2] Sun M, Shariat SF, Cheng C, et al. Prognostic factors and predictive models in renal cell carcinoma: a contemporary review. Eur Urol 2011; 60:644-61.

8 [3] Oudard S, Beuselinck B, Decoene J, Albers P. Sunitinib for the treatment of metastatic renal cell carcinoma. Cancer Treat Rev 2011;37:178-84. [4] Linehan WM, Srinivasan R. Targeted therapies: treating advanced kidney cancer-miles to go before we sleep. Nat Rev Clin Oncol 2013; 10:614-5. [5] Pasic MD, Samaan S, Yousef GM. Genomic medicine: new frontiers and new challenges. Clin Chem 2013;59:158-67. [6] Arsanious A, Bjarnason GA, Yousef GM. From bench to bedside: current and future applications of molecular profiling in renal cell carcinoma. Mol Cancer 2009;8:20. [7] Sanjmyatav J, Steiner T, Wunderlich H, Diegmann J, Gajda M, Junker K. A specific gene expression signature characterizes metastatic potential in clear cell renal cell carcinoma. J Urol 2011;186:289-94. [8] White NM, Yousef GM. Translating molecular signatures of renal cell carcinoma into clinical practice. J Urol 2011;186:9-11. [9] Carlsson L, Nystrom LE, Sundkvist I, Markey F, Lindberg U. Actin polymerizability is influenced by profilin, a low molecular weight protein in non-muscle cells. J Mol Biol 1977;115:465-83. [10] Kwiatkowski DJ, Bruns GA. Human profilin. Molecular cloning, sequence comparison, and chromosomal analysis. J Biol Chem 1988; 263:5910-5. [11] Ding Z, Bae YH, Roy P. Molecular insights on context-specific role of profilin-1 in cell migration. Cell Adh Migr 2012;6:442-9. [12] Witke W, Sutherland JD, Sharpe A, Arai M, Kwiatkowski DJ. Profilin I is essential for cell survival and cell division in early mouse development. Proc Natl Acad Sci U S A 2001;98:3832-6. [13] Wu CH, Fallini C, Ticozzi N, et al. Mutations in the profilin 1 gene cause familial amyotrophic lateral sclerosis. Nature 2012;488:499-503. [14] Joy ME, Vollmer LL, Hulkower K, et al. A high-content, multiplexed screen in human breast cancer cells identifies profilin-1 inducers with anti-migratory activities. PLoS One 2014;9:e88350. [15] Rodriguez FO, Palou J. Predictive factors for recurrence progression and cancer specific survival in high-risk bladder cancer. Curr Opin Urol 2012;22:415-20. [16] Minamida S, Iwamura M, Kodera Y, et al. Profilin 1 overexpression in renal cell carcinoma. Int J Urol 2011;18:63-71. [17] Masui O, White NM, DeSouza LV, et al. Quantitative proteomic analysis in metastatic renal cell carcinoma reveals a unique set of proteins with potential prognostic significance. Mol Cell Proteomics 2013;12:132-44.

J. R. Karamchandani et al. [18] White NM, Newsted DW, Masui O, Romaschin AD, Siu KW, Yousef GM. Identification and validation of dysregulated metabolic pathways in metastatic renal cell carcinoma. Tumour Biol 2014;35:1833-46. [19] Delahunt B, Cheville JC, Martignoni G, et al. The International Society of Urological Pathology (ISUP) grading system for renal cell carcinoma and other prognostic parameters. Am J Surg Pathol 2013;37:1490-504. [20] Budczies J, Klauschen F, Sinn BV, et al. Cutoff Finder: a comprehensive and straightforward Web application enabling rapid biomarker cutoff optimization. PLoS One 2012;7:e51862. [21] Diamandis M, White NM, Yousef GM. Personalized medicine: marking a new epoch in cancer patient management. Mol Cancer Res 2010;8:1175-87. [22] Moch H, Srigley J, Delahunt B, Montironi R, Egevad L, Tan PH. Biomarkers in renal cancer. Virchows Arch 2014;464:359-65. [23] Yousef GM. Personalized cancer genomics: the road map to clinical implementation. Clin Chem 2012;58:661-3. [24] Choudhury Y, Wei X, Chu YH, et al. A multigene assay identifying distinct prognostic subtypes of clear cell renal cell carcinoma with differential response to tyrosine kinase inhibition. Eur Urol 2015;67: 17-20. [25] Gordan JD, Lal P, Dondeti VR, et al. HIF-alpha effects on c-Myc distinguish two subtypes of sporadic VHL-deficient clear cell renal carcinoma. Cancer Cell 2008;14:435-46. [26] Pena-Llopis S, Vega-Rubin-de-Celis S, Liao A, et al. BAP1 loss defines a new class of renal cell carcinoma. Nat Genet 2012;44:751-9. [27] Shuch B, Amin A. Understanding pathologic variants of renal cell carcinoma: distilling therapeutic opportunities from biologic complexity. Eur Urol 2015;67:85-97. [28] Wang L, Williamson SR, Wang M, et al. Molecular subtyping of metastatic renal cell carcinoma: implications for targeted therapy. Mol Cancer 2014;13:39. [29] Mouneimne G, Hansen SD, Selfors LM, et al. Differential remodeling of actin cytoskeleton architecture by profilin isoforms leads to distinct effects on cell migration and invasion. Cancer Cell 2012;22:615-30. [30] Zoidakis J, Makridakis M, Zerefos PG, et al. Profilin 1 is a potential biomarker for bladder cancer aggressiveness. Mol Cell Proteomics 2012;11:M111. [31] Skvortsova I. Profilin 1: do we have a novel proteome-found biomarker predicting response to anticancer therapy? Proteomics 2013;13:2069-71.